The facility layout problem is a combinatorial optimization problem that involves determining the location and shape of various departments within a facility based on inter-department volume and distance measures. An optimal solution to the problem will yield the most efficient layout based on the measures.; The application of Particle Swarm Optimization (PSO) was recently proposed as an approach to solving the facility layout problem. With PSO, potential solutions are produced by dividing departments into swarms of Self-Organizing Tiles (SOT). By following a set of simple behavioral rules based on social information gathered from the environment, the tiles cooperate to produce solutions in a very short amount of time. Initial results provided improvements over CRAFT, one of the primary methods currently used for facility layout.; The main contribution of this thesis work entails evaluating the use of swarm intelligence to produce optimal facility layouts as well as the use of shape measures to assess the quality of produced layouts. The major achievement of this thesis is the design and implementation of a tool that could produce facility layout solutions using Self-Organizing Tiles (SOT). This thesis advances the swarm paradigm by introducing alternative pathways for achieving contiguity of departments.; This thesis utilizes the tool to examine the convergence of SOT on an enumerated optimum for a layout dataset, which requires the exhaustive evaluation of all permutations of a grid layout. The tool was also used to examine the effect of granularity on the ability of SOT to converge on facility layout solutions. A shape metric was utilized as a means of evaluating the quality of produced solutions based on the regularity of the shape of departments, and found that SOT produces fairly regular layouts when granularized to nine tiles per department. Finally, SOT was compared with other algorithms the experimental results revealed that SOT provided minor improvements over currently used methods.
展开▼
